Metadata-Version: 2.4
Name: sciveris
Version: 0.0.0
Summary: Re-executes a study and checks whether the finding is real. Placeholder — public release coming soon.
Project-URL: Homepage, https://github.com/LucZot/sciveris
Project-URL: Source, https://github.com/LucZot/sciveris
Author: Sciveris
License-Expression: Apache-2.0
License-File: LICENSE
Keywords: data-leakage,reproducibility,research-integrity,scientific-computing
Classifier: Development Status :: 1 - Planning
Classifier: Intended Audience :: Science/Research
Classifier: Programming Language :: Python :: 3
Classifier: Topic :: Scientific/Engineering
Requires-Python: >=3.10
Description-Content-Type: text/markdown

# Sciveris

> Reproducibility asks whether the number comes out of the code.
> **Sciveris asks whether the number is _real_.**

**Sciveris** — Latin, _"you will have known."_ (say _shee-VER-is_)

Sciveris re-executes a data-driven analysis, reproduces its headline finding, then adversarially
stress-tests whether that finding holds up — checking for **data leakage**, **batch / site confounds**,
**effect-size inflation**, and **p-hacking** — and issues a **Robustness Report Card**
(`CERTIFIED` · `FLAGGED` · `INCONCLUSIVE` · `ERROR`), each verdict backed by a reproducible exhibit.

It **certifies and strengthens** findings — it doesn't scold.

---

🚧 **This release is a placeholder.** It reserves the name while the first public version is prepared.
It installs nothing useful yet — running `sciveris` just prints where the real tool will land.

Follow along at **https://github.com/LucZot/sciveris**.
